AI Activity Guidelines

Using AI to Learn Data Science

Author

Dr. Cheng-Han Yu

Modified

March 31, 2026

This assignment rewards thinking, judgment, and learning, NOT perfect answers. Strong work shows curiosity, skepticism, and clear reasoning. If you have questions about appropriate AI use, ask before submitting.

Quick Start
  • Work in groups of 3 (or 4).
  • Roles: Prompt Engineer, Data Science Auditor, Synthesizer. See Section 4.
  • Group submits: (1) AI Interaction Log, (2) Human Authored Synthesis, (3) Slides. See Section 6.
  • Each student submits: Individual Reflection (150 to 200 words) See Section 6.
  • Email Dr. Yu all materials by 11:59 PM on the day your team presents. See Section 7.
  • Not allowed: AI text copied into synthesis; undocumented AI use.
  • Grading: rubric categories and total points. See Section 8

In this activity, you will use an AI as a learning partner to explore a data science concept from this course. The goal is NOT to report what AI says. The goal is to learn to:

  • Ask effective questions

  • Evaluate the quality and limitations of AI responses

  • Apply data and modeling reasoning

  • Synthesize understanding in your own words

This activity mirrors how data scientists increasingly use AI in professional practice: as a tool that requires human judgment, verification, and interpretation.

Learning Objectives

By completing this activity, you will be able to:

  1. Use AI strategically to support learning in data science

  2. Critically evaluate AI-generated explanations and claims

  3. Demonstrate conceptual understanding of a data science topic

  4. Communicate statistical reasoning clearly and accurately

  5. Reflect on what human expertise adds beyond AI output

Task Overview

Your group will:

  1. Receive a data science question or concept from the course

  2. Use an AI to help you learn about this topic

  3. Critically evaluate the AI’s responses

  4. Produce a corrected, human-authored explanation

  5. Present your findings to the class

Group Structure

  • Groups of 3 (or 4) students
  • Each group completes one shared investigation
  • Each student has a defined role and submits an individual reflection

Assigned Roles (Rotate Within the Group)

Each group must assign the following roles:

  1. Prompt Engineer

Primary responsibility: Design purposeful prompts with a clear learning goal and document the AI interaction log.

  1. Data Science Auditor

Primary responsibility: Checks AI responses for correctness, assumptions, and limitations.

  1. Synthesizer

Primary responsibility: Convert the group work into a clear human authored explanation and leads the presentation.

[Note:] One team has four members, so two students may share one role. If a role is shared, each student must complete a distinct subtask and clearly label their contributions in their individual reflection. Specific subtask assignments will be provided with your team’s topic.

Approved AI Use

You are expected to use AI for this assignment.

You must:

  • Design your own prompts
  • Document AI interactions
  • Critically evaluate AI responses

You may NOT:

  • Submit AI text verbatim as your own explanation
  • Rely on AI without critique or verification
  • Hide or omit AI usage

Transparency is required.

Required Deliverables

  1. AI Interaction Log (Group Submission)

Submit a short log documenting your AI use.

Include:

  • 3–5 prompts used
  • Selected AI responses
  • A brief annotation after each response addressing:
    • What was the goal of this prompt?
    • What did AI get right?
    • What was incomplete, misleading, or incorrect?

This log is evidence of your learning process.

  1. Human-Authored Synthesis (Group Submission)

Write a clear explanation of the concept in your own words.

Your synthesis must:

  • Be conceptually accurate
  • Improve upon or correct AI output
  • Explain assumptions, limitations, or implications
  • Be understandable to a classmate

AI-generated text should not appear verbatim.

  1. Group Presentation (12–15 minutes)

Your presentation must follow this structure:

  1. The question or concept you investigated
  2. What AI suggested
  3. Where AI fell short or made errors
  4. Your corrected understanding
  5. One key takeaway for future data science work

Slides should be clear, concise, and focused on reasoning rather than volume.

  1. Individual Reflection (Individual Submission)

Each student submits a 150–200 word reflection addressing:

  • What did you personally learn from this activity?
  • What did AI help you understand?
  • What did you need to add, correct, or rethink yourself?
  • How did your role contribute to the group’s work?

This reflection is used to assess individual understanding and contribution.

Submission

  • Submit your AI Interaction Log, Human-Authored Synthesis, and Individual Reflection as a single PDF, or provide a direct link to downloadable PDFs. Indicate who plays which role.
  • Submit your slides in your preferred format, or provide a direct link to view the slides.
  • Email all materials to .
  • Deadline: 11:59 PM on the day your team presents.
  • No late submission of any work is allowed.

Evaluation Criteria

Your group work will be evaluated based on:

  • Quality and intentionality of AI prompts
  • Critical evaluation of AI responses
  • Depth and accuracy of data science understanding
  • Quality of human synthesis beyond AI output
  • Clarity and organization of communication

Individual reflections will be used to adjust individual scores if contributions differ.

Rubric Summary

Criterion Points
AI Prompt Quality 20
AI Documentation 15
Critical Evaluation of AI 25
Data Science Understanding & Synthesis 25
Presentation and Q&A 10
Participation* 5
Total 100

In-Class Participation

  • The presenting group must prepare one or more questions for the other five groups.

  • Each non-presenting group must ask the presenting group at least one question.

  • Each student must submit a brief note describing what they learned from the presentation.

Grading Rubric

Total: 100 points

This rubric is used for all groups, regardless of topic. Grades reflect how you used AI to learn data science, not what AI produced.

AI Prompt Quality and Intentionality (20 points)

Evaluates: Whether AI was used strategically to support learning.

Level Description Points
Excellent Prompts are clear, targeted, and iterative. Each prompt has a specific learning goal, and follow-up prompts show refinement based on prior AI responses. 18–20
Competent Prompts are reasonable and relevant but mostly one-shot or broad. Limited evidence of iteration or strategic refinement. 14–17
Developing Prompts are vague, generic, or copied. Little connection between prompts and learning goals. 8–13
Insufficient Prompts are missing, inappropriate, or unrelated to the assigned topic. 0–7

Documentation of AI Use (AI Interaction Log) (15 points)

Evaluates: Transparency and completeness of AI documentation.

Level Description Points
Excellent AI interaction log is complete, well-organized, and clearly annotated. Goals, strengths, and limitations of AI responses are explicitly documented. 14–15
Competent AI interaction log is mostly complete with basic annotations. Some explanations lack depth or clarity. 11–13
Developing AI interaction log is incomplete or weakly annotated. Limited explanation of AI strengths or limitations. 6–10
Insufficient AI use is poorly documented, unclear, or missing. 0–5

Critical Evaluation of AI Output (25 points)

Evaluates: Ability to question, critique, and challenge AI responses.

Level Description Points
Excellent Clearly identifies specific limitations, assumptions, or errors in AI output and explains why they matter using statistical reasoning. 23–25
Competent Identifies some limitations or weaknesses in AI output, but explanations are partial or underdeveloped. 18–22
Developing Mentions limitations superficially or inconsistently. Limited justification or reasoning. 10–17
Insufficient Accepts AI output uncritically or fails to identify meaningful issues. 0–9

Data Science Understanding and Human Synthesis (30 points)

Evaluates: Depth, accuracy, and originality of the group’s explanation.

Level Description Points
Excellent Explanation is accurate, conceptually clear, and fully human-authored. Demonstrates strong statistical reasoning and clearly improves upon AI output. 27–30
Competent Explanation is mostly accurate and understandable but adds limited insight beyond AI output. Minor gaps or imprecision. 21–26
Developing Explanation shows partial understanding, conceptual gaps, or heavy reliance on AI phrasing. 12–20
Insufficient Explanation is incorrect, unclear, or largely AI-generated with minimal human synthesis. 0–11

Communication and Presentation Quality (10 points)

Evaluates: Clarity and structure of the group presentation.

Level Description Points
Excellent Presentation is well-structured, clear, and focused on reasoning. Effectively explains what AI missed and what the group learned. 9–10
Competent Presentation is understandable but uneven or overly descriptive. Some focus on content rather than reasoning. 7–8
Developing Presentation lacks clarity or structure. Key points are unclear or poorly explained. 4–6
Insufficient Presentation is disorganized, incomplete, or difficult to follow. 0–3

Individual Adjustment (Reflection-Based)

Individual reflections may be used to adjust individual scores up or down if:

  • A student demonstrates substantially stronger or weaker understanding than the group product suggests
  • A student did not meaningfully contribute to the group’s work